The Sciences of the Artificial Herbert A. Simon Prefaces & Chapter 1

The Natural Natural science – Body of knowledge about objects or phenomena in the world – To make the complex simple “to find pattern hidden in apparent chaos”

The World We Live In

Describing the World We Live In The world we live in is filled with physical and conceptual artifacts people have designed to meet specific goals To understand these artifacts we need a way of describing – The natural laws embodied in the artifacts – The human purpose embodied in the artifacts Should there be a science of the artificial? – The title of this book indicates Simon’s opinion

The Artificial The word artificial is somewhat of a pejorative – Here we will use the term as neutrally as possible Many verbs used to describe various forms of creation – Artifice, synthesize, design, compose, … Descriptive vs. normative analyses – Description: an increase in greenhouse gases causes the average temperature to rise – Normative: policies should reduce the emissions of greenhouse gases

First Four Boundaries for Science of the Artificial 1.Artificial things are synthesized (though not always or usually with full forethought) by human beings. 2.Artificial things may imitate appearances in natural things while lacking, in one or many respects, the reality of the latter. 3.Artificial things can be characterized in terms of functions, goals, adaptation. 4.Artificial things are often discussed, particularly when they are being designed, in terms of imperatives as well as descriptives.

Environment as a Mold Fulfillment of purpose involves – Purpose or goal – Character of the artifact – Environment of performance Natural science can describe aspects of the artifact and its environment – Whether a clock can keep time and whether a knife can cut depends on the artifact and where it is being applied

Artifact as Interface We can reason about artifacts as an interface between – the inner environment of the artifact and – the outer environment of its application This can simplify reasoning and enable prediction

Functional Explanations We can predict behavior from knowledge of goals and outer environment with little knowledge of inner workings Only a few characteristics of outer environment may influence success – E.g., Polar animals are likely to blend into a white background for safety and have hair or fat to regulate body temperature Descriptions of designs can avoid providing details of either the outer or inner (e.g. patent descriptions)

Limits of Adaptation Designing is more complicated than just creating a specification – Need to show at least one inner system that meets goals When designs fail we learn more about the relations between the inner and outer environment

Understanding by Simulating An old question – Can a simulation tell us anything we do not already know? Relation to computers and programming People often know things but not foresee their implications – The Moniac

Simulations and Simple Interfaces Skyhook-skyscraper construction of science – Only possible because each level depends on only a very approximate, simplified, abstracted characterization of the system at the next levels. Computers enable simulation based on generic functional descriptions Computer science as an empirical science – Designing reliable systems from unreliable components – Design of time-sharing systems – Sometimes too little is known about task environments and have to just build to evaluate

Computers as Symbol Systems Simon (p. 21) “… if computers are organized somewhat in the image of man …” – Relationship between computer and cognitive science is murky Symbol systems – Entities act as atomic symbols – Symbol structures are expressions about symbols – Symbols can be used to model world but needs Means to acquire information from external environment Means to initiate actions on environment

Closing and Prelude for Following Discussions Hypothesis – “a physical symbol system of the sort I have just described has the necessary and sufficient means for general intelligent action.”